v0.5.2

• updated dependencies
• added support for new gpt4o model

v0.5.1

• Updated support for latest GPT4-Turbo and GPT3Dot5Turbo16K models
• Fixed image generation failing on discord API call

v0.5.0

• Dalle 3 support
• Updated pricing for the latest gpt models
• Updated thread title generation to gpt3.5-turbo-instruct

v0.4.7

• Updated model pricing and token calculation information according to the latest OpenAI update
• Fixed missing token calculation for name field

v0.4.6

• Fixed missing error handling for gpt command followup
• Added context option limit check due to discord character limitation. If context is above limit, just provide it as a file with context-file option
• Bot will now add user to the thread when its created on gpt command. Greatly improves visibility.

v0.4.5

• Added Context file option: upload text file that will be set as a system message. This helps with longer contexts, that were not possible to set in the regular context option due to discord message limits
• Typing indicator will now be sent continuously when bot takes longer to respond. This helps with understanding that request is still going and was not lost
• Result message from /image command had visual update now utilizing discord embeds and using openai generated urls, instead of uploading images manually
• Updated go-openai library to v1.9.3
• Bot should now crash if it failed to register commands, as that feature is the backbone of this application

v0.4.4

• /chat gpt command embeds style improvements
• Added price calculation for GPT 4 32K model

v0.4.3

Added price calculation for gpt 4 8k models

v0.4.2

You can now specify OpenAI ChatCompletion models available for the bot in the configuration file

v0.4.1

• Gpt is now a subcommand of /chat. This adjustment was made to account for potential future expansions of other api/models usage
• Thread will be locked by default if the first OpenAI request failed (e.g. api down, model doesn't exist, etc.)
• Implemented thread token adjustments to avoid reaching token limit of the model. From now on the bot will automatically drop older messages to avoid reaching said limit.